Chief Scientist - Physical AI

About the role

Work at the forefront of Embodied AI and Spatial AI, collaborating with researchers and engineers to solve real‑world challenges and bridge cutting‑edge research with scalable deployment.

Key responsibilities

• Develop state‑of‑the‑art Physical AI approaches, including robotic foundation models, world models, multimodal representation learning, reasoning and planning, as well as imitation and reinforcement learning with integrated low‑level control.
• Translate advanced problem formulations and model designs into efficient, production‑grade systems that enable rapid iteration through large‑scale experimentation.
• Design and execute experiments on both simulation and real‑world robotic platforms, covering data collection, evaluation, and performance analysis.
• Drive closed‑loop learning by integrating data, models, and real‑world feedback to continuously improve system robustness and adaptability.
• Collaborate cross‑functionally across research, infrastructure, and robotics teams to ensure end‑to‑end system coherence and scalability.
• Lead high‑impact research outputs, including patents and publications in top‑tier venues such as NeurIPS, ICML, ICLR, CVPR, and ICRA.

Required profile

• PhD in Computer Science, Robotics, AI, or a related field (or Master’s with equivalent research and industry experience).
• Strong foundation in generative AI and sequence modeling, especially vision‑language or multimodal learning systems.
• Proven research track record with first‑author publications in top‑tier conferences (NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, ICRA).
• Solid expertise in modern deep‑learning frameworks (e.g., PyTorch, TensorFlow) and strong proficiency in Python.
• Hands‑on experience with robotics systems, including decision‑making, planning, and control; experience with real‑world robots or ROS is an asset.
• Experience working with large‑scale datasets, sensor data, and advanced architectures such as transformers or diffusion models.

Required skills

• Python
• PyTorch
• TensorFlow
• ROS
• Large‑scale dataset handling
• Sensor data processing
• Transformers
• Diffusion models